Commit Graph

160 Commits

Author SHA1 Message Date
Hajime Hoshi
bd85d0e08d audio: Add TestPauseBeforeInit
This confirms that Pause before initialization should not launch
another loop.
2019-05-02 21:39:08 +09:00
Hajime Hoshi
7e2a679b17 audio: Context should be reset for each test
As the context remembers the last error, the context should be
re-created for each test.
2019-05-02 21:35:20 +09:00
Hajime Hoshi
396f257629 audio: Fix comments 2019-05-02 04:10:33 +09:00
Hajime Hoshi
876340da17 audio: Bug fix: context.ready should be set true after writing 2019-05-02 03:27:43 +09:00
Hajime Hoshi
2e35d9dc70 audio: Refactoring 2019-05-02 03:21:56 +09:00
Hajime Hoshi
5dd12afd38 audio: Stop the loop when pausing
This change reduces the necessity to fill zero bytes.
2019-05-02 03:16:45 +09:00
Hajime Hoshi
5121b072f1 audio: Change init from a variable to a channel 2019-05-02 02:37:32 +09:00
Hajime Hoshi
bea5b0f685 audio: Add comments 2019-05-02 02:11:35 +09:00
Hajime Hoshi
c112c31e7e audio: Reimplement audio by using multiple oto.Player 2019-04-30 18:20:52 +09:00
Hajime Hoshi
339e76afec audio: Bug fix: Dead lock at seeking
When sending something to channels, a lock should not be used,
or the receiver side might be using the same lock.

Bug: #855
2019-04-30 12:51:22 +09:00
Hajime Hoshi
e19479d243 audio: Add comments about explanation how deadlock happens
This might be related to #855
2019-04-30 01:41:34 +09:00
Hajime Hoshi
5ecc595ce4 audio: Bug fix: Reset playing state at the end of the readLoop 2019-04-29 22:31:10 +09:00
Hajime Hoshi
d2979e288d audio: Bug fix: runningReadLoop must be set to true before starting the goroutine
This can cause deadlock by multiple goroutines if runningReadLoop
is set to true in the goroutine.

Related to #855?
2019-04-29 21:46:26 +09:00
Hajime Hoshi
89f6eba925 Revert "audio: Bug fix: Make ensureReadLoop exactly atomic"
This reverts commit 835497edd3.

Reason: This causes a dead lock at examples/piano with many notes.
This might be related to #855
2019-04-29 14:57:55 +09:00
Hajime Hoshi
fe3e8e376a audio: Move seeking from the loop to Seek function
Seek should work in both cases the loop runs and not. This change
is a preparation.
2019-04-29 02:27:15 +09:00
Hajime Hoshi
c4c6acacf9 audio: Remove (*mux).hasPlayer 2019-04-29 01:01:29 +09:00
Hajime Hoshi
835497edd3 audio: Bug fix: Make ensureReadLoop exactly atomic 2019-04-29 00:05:54 +09:00
Hajime Hoshi
b1607de99d audio: Refactoring 2019-04-28 23:58:06 +09:00
Hajime Hoshi
14b6e95217 audio: Use a boolean value to inidicate inited state intead of channels
This is a preparation to use oto.Players for Players. By using a
boolean value, suspending state can be detected from multiple
oto.Players.
2019-04-28 19:53:28 +09:00
Hajime Hoshi
fe55e55b7e audio: Refactoring 2019-04-28 19:53:28 +09:00
Hajime Hoshi
1327e2239c audio: Use a boolean value to indicate suspending state instead of channels
This is a preparation to use oto.Players for Players. By using a
boolean value, suspending state can be detected from multiple
oto.Players.
2019-04-28 19:53:28 +09:00
Hajime Hoshi
31758bf3ab audio: Remove comment 2019-04-28 19:49:37 +09:00
Hajime Hoshi
2bea7d4e1a audio: Error on duplicated source
Fixes #853
2019-04-28 19:39:46 +09:00
Hajime Hoshi
70a225fd7e audio: Use dummy hooks from tests 2019-04-28 19:35:59 +09:00
Hajime Hoshi
81549a3523 audio: Bug fix: Remove redundant Unlock 2019-04-28 02:44:32 +09:00
Hajime Hoshi
9fc2ae158c audio: Bug fix: Close always returned error 2019-04-28 02:22:11 +09:00
Hajime Hoshi
e4f0a0aa04 audio: Bug fix: Remove depenency on finalizer
This change fixes playerImpl leak by removing dependency on
finalizer usages. Now readLoop can (or should) be called multiple
times even after closing. Only when (*Player).Close is called
explicitly, the read loop cannot be started again.

Fixes #852
2019-04-28 02:03:55 +09:00
Hajime Hoshi
4f831de008 audio: Remove (*playerImpl).sync and use mutex instead
This is a preparation to work syncing wihtout the read-loop.
2019-04-28 01:37:52 +09:00
Hajime Hoshi
acc4ef4e11 audio: Add comments 2019-04-01 01:59:10 +09:00
Hajime Hoshi
ed855d6c6c audio: Refactoring: remove the unused returning value 2019-03-02 04:42:19 +09:00
Hajime Hoshi
a3eddeb50f Improve panic messages 2019-02-06 23:43:03 -10:00
Hajime Hoshi
a4123a479e audio: Initialize the context first
This is necessary to create an audio's Player before the loop if
we want to create *oto.Player for each *audio.Player.
2019-01-30 00:45:36 +09:00
Hajime Hoshi
13b0d82b6d audio: Use oto.NewContext 2019-01-30 00:36:47 +09:00
Hajime Hoshi
6d714a16cf audio: Remove unneeded 'break' 2019-01-10 02:34:58 +09:00
Hajime Hoshi
a6b3f761f8 audio: Refactoring: Remove dependency on clock package 2019-01-10 00:59:38 +09:00
Hajime Hoshi
14404334e7 audio: Rename players to mux 2019-01-09 22:10:52 +09:00
Hajime Hoshi
1760070130 audio: Add comments about GC (#746) 2018-12-18 23:07:19 +09:00
Hajime Hoshi
b9eba54115 audio: Use a dummy driver for testing
Fixes #750

This also fixes an out-of-index bug when all players are skipped.
2018-12-17 23:06:59 +09:00
Hajime Hoshi
b0cb216f5f audio: Close audio after its playing finishes
Fixes #746

This is a temporal fix and we will need further re-designing of
audio package.
2018-12-17 04:02:32 +09:00
Hajime Hoshi
08a369b8fd audio: Bug fix: GC audio players correctly
Fixes #746
2018-12-15 22:33:45 +09:00
Hajime Hoshi
d14202ef7b audio: Add Context.IsReady
Fixes #717
2018-10-16 00:02:19 +09:00
Hajime Hoshi
2969de5b1f audio: Refactoring: Redefine bytesPerSample 2018-07-24 01:54:01 +09:00
Hajime Hoshi
d752408da3 Fix comments not to use the term 'this function' 2018-07-15 01:04:46 +09:00
Hajime Hoshi
273093b237 audio: Skip the player in the state of starting, seeking or EOF
This is basically reland of
2fee7a6fe5.

Before this change, if a player's buffer was not enough for
reading, 0 value were used and this caused noises. The reading
size should be aligned with all the players.

However, there are some cases that the player should be skippped.
For example, just after a player just starts playing or seeking,
the buffer is empty. In this case, other players should not wait
for the player since decoding might take some time. Another case
is that the player reached EOF.

This change aligns the read buffer sizes but use zero values only
when the player just starts or seeks, or reaches EOF.
2018-06-09 22:00:11 +09:00
Hajime Hoshi
facf184548 Revert "audio: Consider the state of the player's starting and seeking"
This reverts commit 2fee7a6fe5.

Reason: This causes freezing on browsers
2018-06-09 19:28:20 +09:00
Hajime Hoshi
b62e07b9bd audio: Remove println 2018-06-09 19:05:07 +09:00
Hajime Hoshi
2fee7a6fe5 audio: Consider the state of the player's starting and seeking
Before this change, if a player's buffer was not enough for
reading, 0 value were used and this caused noises. The reading
size should be aligned with all the players.

Just after a player just starts playing or seeking, the buffer is
empty but other players should not wait for the player read since
decoding might take some time.

To summerize, this change aligns the read buffer sizes but use
zero values only when the player just starts or seeks.
2018-06-09 19:01:57 +09:00
Hajime Hoshi
91958d24ef audio: Use bigger buffers for mobiles 2018-06-09 18:14:18 +09:00
Hajime Hoshi
faaf391619 audio: Give up syncing audio timer and game timer (#617)
Syncing was already incomplete (e.g. decoding takes more than one
frame and delays can happen in this case). Giving up syncing audio
timer and game timer should not affect the game experience so much.
Instead, clock implementation will be much simpler.
2018-05-27 00:30:12 +09:00
Hajime Hoshi
5976e4bbbc audio: More intellegent suspending/resuming (#617)
Before this change, the audio is suspended when the game stops for
1/12[s]. However, as game often stops for more than 1/12[s]
especially on mobiles, this implemntation caused some audio
glitches.

This change fixes this problem by re-implementing suspending/
resumeing audio by detecting the window is active/focused or not.
2018-05-26 23:31:04 +09:00